Etiology
Etiology of Watermelon Rind Necrosis
D. L. Hopkins and G. W. Elmstrom
Associate Professor of Plant Pathology and Associate Professor of Horticulture, respectively, University of Florida,
Agricultural Research Center, Leesburg, FL 32748.
Florida Agricultural Experiment Station Journal Series Paper No. 109.
Accepted for publication 2 March 1977.
ABSTRACT
HOPKINS, D. L., and G. W. ELMSTROM. 1977. Etiology of watermelon rind necrosis. Phytopathology 67: 961-964.
Bacteria were isolated consistently from healthy
watermelons and from melons with watermelon rind necrosis
(WRN). No single bacterial species was isolated consistently
from WRN tissue. The diversity of the internal bacterial flora
of melons with WRN was similar to that of healthy melons,
However, enterobacteria were isolated more frequently from
melons with WRN than from healthy melons. Most of the
bacteria isolated produced localized rind necrosis when
injected at high concentrations but at low concentrations
some bacteria were more effective than others. With some
isolates a systemic rind necrosis was produced which
resembled that often observed in natural infections. In field
studies, foliar sprays of mineral nutrients did not affect WRN
incidence. It is hypothesized that the WRN symptoms result
from the multiplication of one of the resident bacteria to a
level high enough to induce a necrotic reaction in the
watermelon.
In Georgia, an internal browning of watermelon was
attributed to prolonged drought conditions during fruit
maturation in 1925 (5). A relatively inconspicuous
vascular browning of the rind was observed in Hawaii and
attributed to physiological causes (11). Bacterial rind
necrosis of watermelon also has been reported from
Hawaii (11), Texas (18), Florida (8), and California (13).
Symptoms of watermelon rind necrosis (WRN) include a
brown, dry, and hard necrosis of the rind that rarely
extends into the flesh of most watermelon cultivars. The
only external symptom is misshapen fruit in a few severely
affected melons. Severely affected melons are
unmarketable when sliced,
Significant differences have been observed in the
incidence and severity of rind necrosis in various
watermelon cultivars. Round-fruited cultivars tend to
have a higher incidence of WRN than long-fruited ones.
Incidence of WRN is high enough in some cultivars to
make them unsuited for production in Florida (3, 9).
Watermelon rind necrosis was reported to be caused by
a species of Erwinia resembling E. carnegieana(11, 18). In
our preliminary studies, bacteria consistently were
isolated from the necrotic areas, and Erwinia-type
bacteria were most common. However, the localized
necrosis symptoms of WRN were reproduced at the
inoculation site when any of several bacterial types were
injected into the rinds. Extensive systemic browning of
the rind that is often found in natural infections did not
occur (8). Symptoms of WRN resemble those that have
been reported for boron and calcium deficiencies in
various fruits and vegetables. Our objective was to
determine whether or not the watermelon rind necrosis
disease is caused by a single bacterium and to evaluate the
role of nutrition, host resistance, and other factors in the
Copyright © 1977 The American Phytopathological Society, 3340
Pilot Knob Road, St. Paul, MN 55121. All rights reserved,
disease syndrome. Preliminary reports have been made
on parts of this study (4, 8, 10).
MATERIALS AND METHODS
Isolations from watermelon rinds were made by peeling
off the surface of the rind with a sterile scalpel and
comminuting a small internal section (8-15 mm3 ) of
necrotic rind in 1 ml of sterile water. This suspension was
streaked on plates of King's Medium B (12) and, after an
appropriate period of incubation, stock cultures were
prepared from single colonies either directly or after
dilution plating. Stock cultures were maintained in
sterile, demineralized water at room temperature.
Standard bacteriological tests, conducted as described
elsewhere (1, 2, 14), were used to identify the various
bacterial types which were isolated from watermelons.
Isolates were considered to be members of the Enterobacteriaceae if they were Gram-negative rods,
facultatively anaerobic, and either motile by peritrichous
flagella or nonmotile. Enterobacteria which grew as
yellow colonies on nutrient agar, produced little or no gas
from glucose, and did not decarboxylate arginine, lysine,
or ornithine were classified as Erwinia herbicola types.
The fluorescent pseudomonads were isolates which were
Gram-negative, aerobic, motile by polar flagella, and
produced diffusible yellow-green fluorescent pigments on
King's Medium B (12). Isolates were classified as
xanthomonads if they were Gram-negative, aerobic,
motile by a polar flagellum, and produced a non-watersoluble yellow pigment on nutrient agar. All
bacteriological tests were performed at least twice on each
isolate. Known isolates were included as standards in the
tests.
Watermelon fruits were inoculated by injecting
bacterial suspensions into the rind. The inoculum was
961
962
PHYTOPATHOLOGY
adjusted to 108 viable cells/ ml by measuring OD600nm with
a Spectronic 20 colorimeter, and other inoculum
concentrations were prepared by dilution of this adjusted
suspension. Inoculum suspensions (0.5 ml) were injected
into each of three sites per melon. Fruits which were at
least 2 wk from maturity were used for inoculations. At
maturity, melons were first sliced through the center of
the inoculation sites and then sliced into 5 cm wide, or
smaller, slices. Localized and systemic symptoms were
recorded.
To determine rates of multiplication of bacterial
isolates in watermelon furit, one-third to full-sized
immature melons were inoculated by injecting 0.5 ml of
bacterial suspension, containing either 10' or 10' viable
cells/ml, at three locations on the fruit. Inoculated
melons either were stored at 26 ± 5 C or left attached to
the vine in the field. Melons were sampled periodically to
determine bacterial concentrations at inoculation sites.
The surface of the rind was cut from the injected spot with
a sterile scalpel and a disk of tissue was removed from the
inoculated area with a sterile 6-mm-diameter cork borer.
Aseptically, a 1-mm slice was cut transversely from the
disk of tissue and placed in 2 ml of sterile water. The slice
TABLE I. Frequency of isolation of various bacterial types
from watermelons with and without rind necrosis
_
Rind necrosisa
Healthy'
Bacterial type
1974
1975
1975
Enterobacteriaceae:
61
54
23
Erwinia herbicola
39
12
9
Other enterobacteria
32
42
18
Fluorescent pseudomonads
8
8
5
Xanthomonads
10
8
0
Misc. Gram-negative
8
35
68
Gram-positive
26
50
36
aData are given as percentage of melons from which the
bacterial type was isolated. More than one type was isolated
from some melons. Isolations were obtained from 51 melons in
1974, 26 melons with rind necrosis in 1975, and 22 symptomless
melons in 1975.
[Vol. 67
was crushed with a sterile glass rod and the bacterial
concentration per milliliter of suspension was determined
by dilution plating. There were at least three replications
of each treatment. Bacterial colonies were counted 24-48
hr after plating.
Field experiments were conducted to determine the
effect of supplemental boron and calcium on the
incidence of WRN. In 1973 the cultivars Charleston Gray
(moderately tolerant to WRN) and Summerfield (highly
susceptible) were planted. In 1974 the test was repeated
with the cultivars Charleston Gray and Klondike Blue
Ribbon, highly susceptible to WRN. Test plots consisted
of 10 hills in two rows with 1.5 m between hills and 3 m
between rows. Each treatment was replicated four times.
Boron (0.24 g/ liter) and calcium (3.1 g/ liter) foliar sprays
were applied to runoff five times at 2-wk intervals
beginning 13 March in 1973 and 28 March in 1974.
Mature fruits were harvested between 28 May and 3 July.
All melons were sliced in the field and WRN incidence
was recorded.
RESULTS
Isolations from melons with watermelon rind
necrosis.-In 1974, bacteria were isolated from 126 of 132
melons with WRN and 25 of 30 symptomless ones. In
1975, bacteria were isolated from 29 of 30 WRN melons
and 27 of 30 symptomless melons.
Bacterial isolates obtained from 51 WRN melons in
1974 and 26 WRN melons in 1975 were characterized in
an attempt to determine whether any one bacterium or
bacterial type could be associated consistently with the
disease (Table 1). Enterobacteria were the most
frequently isolated bacterial type. However, in 1974,
more than half the enterobacteria were Erwiniaherbicola
types, which are common on plant surfaces and as
secondary invaders. Gram-positive bacterial types also
were common,
but Pseudomonads,
these consisted of
at least four or and
five
different
bacteria.
xanthomonads,
other Gram-negative bacteria also were found in melons
with WRN. No single bacterium could be isolated
consistently.
TABLE 2. Effect of inoculum concentration on watermelon rind necrosis symptom development in two watermelon cultivars after
bacterial injection
Infectivity' of various
inoculum concentrations
(cells/ml)
Cultivar and
isolate
Sweet Princess
(tolerant)
RN71-3
RN74-10
RN74-3
Systemic
101
101
101
necrosisb
3/3
9/9
3/6
3/3
5/6
2/6
3/5
1/6
0/4
0/5
5/5
Klondike Blue Ribbon
(highly susceptible)
RN71-3
9/9
6/6
7/9
5/8
RN74-10
...
...
2/8
0/3
RN74-3
9/9
8/8
8/8
3/9
aData are expressed as number of necrotic sites over the total number of injected sites. There were three inoculated sites per melon.
hData are expressed as number of melons with systemic necrosis over the total number of
melons injected. This includes melons
injected with all three inoculum concentrations. "..." indicates that data was not obtained for that inoculum concentration.
HOPKINS AND ELMSTROM: WATERMELON RIND NECROSIS
August 1977]
Since bacteria could be isolated from most
watermelons, comparisons were made between the
bacterial flora of melons with rind necrosis and that of
symptomless melons (Table 1). The most interesting
difference was the much more frequent occurrence of
enterobacteria in the necrotic tissue than in the
symptomless melons. This difference primarily was
traceable to enterobacteria other than E. herbicola.Based
on the arginine, lysine, and ornithine decarboxylase
reactions and on gas production in glucose (1, 14), these
appeared to be mostly an Enterobacter sp. and a few
Klebsiella and Erwiniatypes. Gram-positive bacteria also
occurred slightly more often in the WRN melons. The
miscellaneous unidentified Gram-negative bacteria were
isolated twice as frequently from symptomless melons as
from melons with WRN.
Pathogenicity tests.-In preliminary tests, WRN
symptoms were reproduced readily at the inoculation
sites by injection with both Erwinia and Pseudomonas
isolates, but the extensive systemic necrosis of the rind
that is found often in natural infections was not observed,
The effect of inoculum concentration on symptom
development was studied in two cultivars with different
levels of susceptibility to WRN (Table 2). Inoculations
were made in the field. All three bacterial isolates
reproduced the necrotic symptom at the site of injection
with inoculum concentrations of only 10' cells/ ml. Even
lower concentrations occasionally gave localized
necrosis. Isolates RN 71-3 (an Enterobactersp.) and RN
74-3 (a pseudomonad) were more effective than was RN
74-10 (an Erwinia herbicola type) in causing necrosis,
109108
-sprays
0combined,
I
"•
6
10O
Geliminate
10 5
0'4//
'0
/
10 3
0
0-
10 2
Fruit rotting bacterium
WRN bacterium
1O7 cells per
per ml
inoculum
105cells
ml inoculum
105 cells per ml inoculum
-
963
Frequency ofnecrosis at the injection site was similar with
the tolerant Sweet Princess and the highly susceptible
Klondike Blue Ribbon.
In contrast to earlier tests, systemic necrosis of the rind
was observed in these inoculated melons (Table 2).
Development of systemic necrosis after inoculation by
injection was much more frequent in Klondike Blue
Ribbon than in Sweet Princess. Systemic necrosis
developed in 1 of 15 inoculated Sweet Princess melons
and in 8 of 20 inoculated Klondike Blue Ribbon melons.
The bacterial isolates also seemed to vary in ability to
cause systemic symptoms. Isolate RN 71-3 was the most
effective and isolate RN 74-10 was ineffective.
Multiplication patterns of watermelon rind necrosis
bacteria versus fruit-rotting bacteria.-The
multiplication pattern in Charleston Gray of WRN
isolate RN 71-3 was compared with that of isolate 64-3
[an isolate of Pseudomonas lachrymans (E. F. Sm. &
Bryan) Carsner] which rots the fruit (Fig. 1). At the lower
inoculum level (10' cells/ ml) both bacteria multiplied at a
similar rate, with the WRN isolate reaching a maximum
between 24 and 48 hr and the fruit-rotting isolate reaching
its maximum 48-72 hr after injection. Both isolates
reached maximum population levels approximately 24 hr
earlier following injection at the higher (10' cells/ml)
inoculum concentration. Necrosis first occurred with the
WRN bacterium when its maximum population was
reached and rotting also occurred when the P.
lachrymans isolate reached its maximum population.
Final populations of the WRN isolate were 20-100 times
less than the final populations of the fruit-rotting isolate.
The rot from the fruit-rotting bacterium spread
throughout the melon, whereas the necrosis from the
WRN bacterium was restricted to the rind. This
experiment was repeated three times. Three other WRN
isolates were found to have similar multiplication
patterns in other tests.
Supplemental boron and calcium.-Foliar nutrient
of boron and calcium, applied both separately and
failed to significantly reduce WRN. The
incidence of WRN was slightly lower in the boron and
calcium spray treatments in 1973, but was higher in 1974.
In no instance did supplemental mineral nutrients
WRN.
Watermelon mosaic virus.-It has been reported that
WRN is associated with watermelon mosaic virus
(WMV) infection (18). However, we observed that
incidence of WRN was not correlated with WM4V
incidence. In our 1973 tests watermelon mosaic was very
severe,
but in 1974 WMV infection occurred rarely if at all
in our test
plots. In contrast, there was a slightly higher
overall incidence of WRN in 1974 than in 1973.
DISCUSSION
0
24 48 72
96 120 144 168
Hours
A mixed bacterial flora commonly is present inside
healthy vegetables, but little is known about the
significance of this flora (7, 15, 16). Of the cucurbits,
Fig. 1. Multiplication of watermelon rind necrosis (WRN)
isolate, RN71-3, and a fruit rotting Pseudomonas lachrymans
isolate at two inoculum levels in Charleston Gray watermelon
cucumbers are known to contain a diverse bacterial flora
(7, 15). In this study we consistently isolated bacteria from
healthy watermelons and from melons with WRN. Thus,
tissue. Bacterial concentration is given as viable cells per
3
milliliter, and each ml of extract represents 14 mm of tissue.
it appears that watermelons also harbor resident bacterial
populations internally.
964
[Vol. 67
PHYTOPATHOLOGY
Systemic necrosis of the rind similar to that often
observed in natural infections was reproduced in these
tests. Frequency of systemic necrosis development
seemed to depend both on susceptibility of the
watermelon cultivar and on the virulence of the
bacterium. This demonstrated that WRN symptoms can
be reproduced totally by bacteria, but did not prove that
they are the primary incitants in natural infections.
We have not observed any nutritional effect on the
disease. The basic fertilizer mix that was used routinely
each year contained fritted trace elements which should
supply adequate amounts of trace elements, still WRN
occurs every year. The failure to reduce WRN with
supplemental mineral nutrients indicates that the disease
is not simply a deficiency disorder in which the bacteria
are secondary invaders.
From these studies, we have concluded that in Florida
WRN may be incited by more than one member of the
normal internal bacterial flora of watermelon. The WRN
disease syndrome appears to result from a general
hypersensitivity.
resistance
reaction ethat
resocistancereationofm
that may
my be
baterl
h ypere
ity. The
the
disease and the multiplication patterns of the incitant
bacteria are consistent with this hypothesis.
With this hypersensitivity hypothesis, there are several
events that must occur for the disease to develop and the
probabilities of these events occurring determines the
susceptibility of a watermelon cultivar. First, there must
be an internal bacterial flora. Although this was
demonstrated in healthy melons, there may be differences
in populations among cultivars. Second, the bacteria
must multiply to a level that initiates the hypersensitive
reaction. In most melons the resident bacteria never
multiply to levels that cause WRN. Perhaps the resident
bacteria are better able to multiply and cause WRN under
certain environmental conditions or at times of stress on
the plant, such as that of WMV infection or nutritional
imbalance. Third, the effectiveness of the hypersensitive
responsedetermines the amount of systemic necrosis that
refore, theamount of temiseas that
response
occurs and, therefore, the severity of the disease. Invery
susceptible watermelon cultivars, such as Klondike Blue
Ribbon, the necrosis seems to spread through the
vascular tissue and results in systemic necrosis around the
rind and even extending into the flesh of the melon.
Perhaps in more tolerant cultivars, the hypersensitive
reaction is more effective in localizing the bacteria,
In many ways WRN is very similar to graywall of
tomato. Both diseases apparently are incited by more
than one bacterium (17), both show cultivar differences in
susceptibility (6), and both exhibit multiplication
patterns of the incitant bacteria that indicate hypersensitivity (17). Rind necrosis of watermelon, graywall of
tomato, and perhaps others may represent a class of
diseases that are incited by bacteria that are normally
residents of the healthy host. When the proper
predisposing environmental conditions are present, these
residents may cause disease.
LITERATURE CITED
1. BUCHANAN, R. E., and N. E. GIBBONS, eds. 1974.
Bergey's manual of determinative bacteriology, 8th ed.
Williams and Wilkins, Baltimore, Maryland. 1246 p.
2. COLLINS, C. H. 1967. Microbiological methods, 2nd ed.
Plenum Press, New York, NY 404 p.
~
3. ELMSTROM, G. W., and D. L. HOPKINS. 1973. Variable
susceptibility to bacterial rind necrosis in watermelon.
HortScience 8:32.
4. ELMSTROM, G. W., and D. L. HOPKINS. 1975.
Watermelon rind necrosis: effect of supplemental boron
and calcium. HortScience 10:145 (Abstr.).
5. GILBERT, W. W., and E. ARTSCHWAGER. 1925.
Watermelon internal browning. Phytopathology 15:119121.
6. HALL, C. B., and R. E. STALL. 1967. Graywall-like
symptoms produced in tomato fruits by bacteria. J. Am.
Soc. Hortic. Sci. 91:573-578.
7. HAYWARD, A. C. 1974. Latent infections by bacteria.
Annu. Rev. Phytopathol. 12:87-97.
8. HOPKINS, D. L. 1972. Association of bacteria with rind
necrosis of watermelon in Florida. Phytopathology
62:804 (Abstr.).
9. HOPKINS, D. L., and G. W. ELMSTROM. 1974. Severity
of
bacterial
in watermelon
cultivars in
Florida.
Proc.rind
Fla.necrosis
State Hortic.
Soc. 87:184-187.
10. HOPKINS, D. L., and G. W. ELMSTROM. 1975. The
internal bacterial flora of watermelon with rind necrosis.
Proc. Am. Phytopathol. Soc. 2:134 (Abstr.).
11. ISHII, M., and M. ARAGAKI. 1960. Bacterial rind necrosis
of watermelon. Plant Dis. Rep. 44:761-763.
12. KING, E. 0., M. K. WARD, and D. E. RANEY. 1954. Two
simple media for the demonstration of pyocyanin and
fluorescin. J. Lab. Clin. Med. 44:301-307.
13. KONTAXIS, D. G., and T. KURUPAS. 1975. Watermelon
rind necrosis in Imperial Valley. Calif. Agric. 29:14-15.
14. LENNETTE, E. H., E. H. SPAULDING, and J. P.
TRUANT, eds. 1974. Manual of clinical microbiology,
2nd ed. American Society for Microbiology,
Washington, D.C. 970 p.
15. MENELEY, J. C., and M. E. STANGHELLINI. 1974.
Detection of enteric bacteria within locular tissue of
healthy cucumbers. J. Food Sci. 39:1267-1268.
16. SAMISH, Z., R. ETINGER-TULCZYNSKA, and M.
BICK.
1963. TheodSi
microflora
within the tissue of fruits and
vgtbe.J
82926
vegetables. J. Food Sci. 28:259-266.
17.
1.STALL, R. E., and C. B. HALL. 1969. Association of
bacteria with graywall of tomato. Phytopathology
59:1650-1653.
18. THOMAS, C. E. 1968. Bacterial rind necrosis of watermelon
in South Texas. Plant Dis. Rep. 52:375-377.